Hi! I wanted to express my appreciation for your team's excellent work. Recently, while conducting experiments on "exp1 debiasing gender", I calculated a gender bias of "0.0546" using my own bias calculation code, which notably outperformed the "0.23±0.16" reported in your paper. This discrepancy has prompted concerns about the correctness of my implementation. Could you please consider releasing your bias calculation code for comparison and validation? It'd help me compare and verify my results. Looking forward to hearing from you. Thanks!!😊
Kqp1227
commented
1 month ago
Hi! I wanted to express my appreciation for your team's excellent work. Recently, while conducting experiments on "exp1 debiasing gender", I calculated a gender bias of "0.0546" using my own bias calculation code, which notably outperformed the "0.23±0.16" reported in your paper. This discrepancy has prompted concerns about the correctness of my implementation. Could you please consider releasing your bias calculation code for comparison and validation? It'd help me compare and verify my results. Looking forward to hearing from you. Thanks!!😊